Change of IO size possible?

[Andreas Bechtle_1]

Hello,

On my servers all datas are on a SAN (RAID5). As I have a lot of IO the bottleneck is on the SAN.
My test:
RAID5 with 7 disk (100 IO/sec each). I use LVM/vxfs and get 30Mb/sec for write-only. That should be 64k/IO.
If I would be able to increase the IO-Size I should get better performance, right? Is it possible to change this value in any way and still work with LVM/vxfs?

Andreas

vxtunefs /amb:
Filesystem i/o parameters for /amb
read_pref_io = 65536
read_nstream = 1
read_unit_io = 65536
write_pref_io = 65536
write_nstream = 1
write_unit_io = 65536
pref_strength = 10
buf_breakup_size = 131072
discovered_direct_iosz = 262144
max_direct_iosz = 1048576
default_indir_size = 8192
qio_cache_enable = 0
write_throttle = 0
max_diskq = 1048576
initial_extent_size = 1
max_seqio_extent_size = 2048
max_buf_data_size = 8192
hsm_write_prealloc = 0
read_ahead = 1

HP-UX11.23/OnlineJFS